https://doi.org/10.1007/s10549-019-05351-4
REVIEW
Assessing the quality and communicative aspects of patient decision
aids for early‑stage breast cancer treatment: a systematic review
Ruben Vromans1 · Kim Tenfelde1 · Steffen Pauws1,2 · Mies van Eenbergen3 · Ingeborg Mares‑Engelberts4,5 ·
Galina Velikova6 · Lonneke van de Poll‑Franse3,7,8 · Emiel Krahmer1 Received: 28 May 2019 / Accepted: 5 July 2019
© The Author(s) 2019
Abstract
Purpose Decision aids (DAs) support patients in shared decision-making by providing balanced evidence-based treatment information and eliciting patients’ preferences. The purpose of this systematic review was to assess the quality and com-municative aspects of DAs for women diagnosed with early-stage breast cancer.
Methods Twenty-one currently available patient DAs were identified through both published literature (MEDLINE, Embase, CINAHL, CENTRAL, and PsycINFO) and online sources. The DAs were reviewed for their quality by using the International Patient Decision Aid Standards (IPDAS) checklist, and subsequently assessed to what extent they paid attention to various communicative aspects, including (i) information presentation, (ii) personalization, (iii) interaction, (iv) information control, (v) accessibility, (vi) suitability, and (vii) source of information.
Results The quality of the DAs varied substantially, with many failing to comply with all components of the IPDAS criteria (mean IPDAS score = 64%, range 31–92%). Five aids (24%) did not include any probability information, 10 (48%) presented multimodal descriptions of outcome probabilities (combining words, numbers, and visual aids), and only 2 (10%) provided personalized treatment outcomes based on patients and tumor characteristics. About half (12; 57%) used interaction methods for eliciting patients’ preferences, 16 (76%) were too lengthy, and 5 (24%) were not fully accessible.
Conclusions In addition to the limited adherence to the IPDAS checklist, our findings suggest that communicative aspects receive even less attention. Future patient DA developments for breast cancer treatment should include communicative aspects that could influence the uptake of DAs in daily clinical practice.
Keywords Breast cancer · Decision aids · Patient education · Risk communication · Shared decision-making · Treatment decision-making
Abbreviations
DA Decision aid
IPDAS International patient decision aids standards SDM Shared decision-making
Introduction
In early breast cancer care, there has been rapid growth in the development of patient decision aids (DAs) to support the process of shared decision-making (SDM) between patients and their clinician [1]. DAs are tools (aimed at patients and distributed by clinicians) that provide information about treatment options and associated risks of side-effects and disease recurrence, and help patients clarify their values and preferences [2, 3]. Moreover, DAs should encourage patients to (actively) participate in the SDM process with their clinician [3, 4]. Despite great promise and the increas-ing interest in developincreas-ing DAs [1, 2], the extent to which they are implemented into daily clinical practice appears to be limited [5, 6].
One reason for this might be the variability in the characteristics and quality of DAs for early breast cancer
Electronic supplementary material The online version of this
article (https ://doi.org/10.1007/s1054 9-019-05351 -4) contains supplementary material, which is available to authorized users. * Ruben Vromans
r.d.vromans@uvt.nl
treatment [7]. Assessing the quality of DAs (e.g., whether the DAs’ content is reliable and evidence-based, or how they were developed and field-tested) is relevant to patients and clinicians [8], since a lack of trust in or familiarity with the quality of DAs could explain why clinicians do not distribute them to their patients [9]. Typically, the validated international patient decision aids standards (IPDAS) checklist is used to ensure the quality of DAs [10], and covers a variety of dimensions, ranging from information about treatment options and outcome prob-abilities to decision guidance and development process. Although the IPDAS is considered the gold standard for developing and evaluating DAs [11], being IPDAS com-pliant does not guarantee that DAs will reach the hands of patients.
We argue that another factor is the extent to which DAs pay attention to the communicative aspects. In fact, DAs include many communication aspects that may influence the use and understanding of the tools by patients and cli-nicians, but are not covered by the IPDAS checklist [12]. These include, for instance, how DAs present information about treatment options and associated outcome prob-abilities to patients (e.g., only words or numbers, or in combination with visual aids) [13], or how they communi-cate uncertainty around statistics. Another communicative aspect is how DAs interact with patients to elicit their val-ues or preferences (e.g., value-clarification exercise) [14], or to provide patients with personalized information based on their personal and tumor characteristics (e.g., personal-ized risk or survival estimates), all of which can improve patient and clinician’s understanding of the personal and clinical situation at hand. Furthermore, aspects like the
suitability (e.g., complex language use), accessibility (e.g.,
only internet-based), or source of information (e.g., reli-able outcome probabilities) could disturb the communica-tion process between the DA, patient, and clinician [15]. All these aspects are important elements of the commu-nication process [16], and DAs that pay less attention to these aspects may limit their ability to be distributed by clinicians and to be used and/or comprehended by patients.
Although some reviews have shown the effectiveness of DAs in early breast cancer care [1, 17, 18], there has been no review on the quality and use of communicative aspects among existing DAs for patients facing early breast cancer treatment decisions. Therefore, the aims of this systematic review were (1) to make an inventory of currently avail-able patient DAs for early-stage breast cancer treatment in both English and Dutch, (2) to critically review their quality based on the IPDAS criteria, and (3) to assess to what extent they pay attention to various communicative aspects.
Methods
This systematic review is conducted and reported in com-pliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [19].
Data sources and search strategy
A systematic search of both published literature and online sources was conducted to identify and obtain DAs for patients facing early breast cancer treatment decisions. To obtain DAs with associated studies through published
literature, we searched the following databases:
MED-LINE (via PubMed), EMBASE, Cochrane Library, The Cumulative Index to Nursing and Allied Health Literature (CINAHL), and PsycINFO. Given that the IPDAS checklist was launched in 2006, we searched the databases from Janu-ary 2006 until March 2018. Reference lists and author names were searched to identify additional publications that met the eligibility criteria. The search strategy included a combina-tion of keywords, synonyms, and MeSH headings relating to the concepts of breast cancer, DAs, SDM, and treatments (Supplementary Material 1). To obtain DAs without associ-ated studies through online sources, we searched the Ottowa Decision Aid Library Inventory (https ://decis ionai d.ohri.ca/ cochi nvent .php), and GoogleTM (search terms “decision aid,” “breast cancer,” and “treatment”) in both Dutch and English for which the first 100 hits were analyzed.
Inclusion and exclusion criteria
We developed inclusion and exclusion criteria for the iden-tification of scientific studies and for decision aids. For the
studies obtained through published literature, the inclusion
criteria include those that were (1) reported in a scientific journal (peer-reviewed); (2) published between 2006 and 2018; (3) written in English or Dutch. Study types eligible for inclusion were (1) (non-)randomized controlled trials or experimental studies that addressed the impact of DAs as intervention on decisional outcomes or treatment choice; (2) development and/or evaluation of the DAs (e.g., protocol, developmental, evaluation, usability testing, or observational studies). Target populations of studies included newly diag-nosed patients with early-stage breast cancer facing treat-ment decision-making.
For both DAs obtained through published literature and online sources, the following exclusion criteria applied: DAs (1) developed for women with advanced stages of breast cancer or for breast cancer screening; (2) in the for-mat of predictive or decision-support tools (e.g., Predict-UK, Adjuvant!Online) since such tools are aimed for both
clinician and patients; (3) in the format of phone calls, online support groups, interviews, nomograms, or audiotapes, since such formats could not be analyzed. Finally, the following inclusion criteria applied: DAs that were (1) published between 2006 and 2018; (2) (publicly) available; (3) fully accessible (e.g., no monetary costs associated with the DA such as one time purchase, or no need to be prescribed by a certain healthcare system or clinician); (4) written in English or Dutch.
Study and decision aid selection
Two reviewers (RV, KT) screened all retrieved articles for relevance based on title and abstract for initial eligibility. The overall kappa score for inter-rate agreement during the screening phase was strong (κ = 0.97) [20]. Afterwards, the few disagreements were resolved through discussion or adjudication by a third person. Subsequently, the same two reviewers independently evaluated the articles that passed the previous screening phase based on the eligibility crite-ria and disagreements were resolved through discussion and consensus between the two reviewers. The overall kappa score during the study eligibility phase was strong (κ = 0.91). Data extraction of the included studies and DAs were inde-pendently assessed by two reviewers.
Assessment of decision aids
The assessment of the identified DAs consisted of two parts. DAs were first reviewed for their quality according to IPDAS criteria, after which they were critically assessed on a com-municative aspect checklist. Each DA was independently assessed by two coders (four coding teams in total). Inter-rate agreements (κ) achieved by the teams ranged from 0.74 to 0.86 for the IPDAS checklist (mean κ = 0.81), and from 0.76 to 0.90 for the assessment of CAs (mean κ = 0.83). The total, average inter-rate agreement was good (κ = 0.82).
Quality of decision aids
Quality of the included DAs was assessed by using the IPDAS Collaboration criteria framework. The IPDAS instrument (Supplementary Material 2) [10] consists of 36 items divided into eight dimensions: (i) information about
options (items 1–8), (ii) outcome probabilities (items
9–16), (iii) clarifying values (items 17–20), (iv) decision
guidance (items (21–22), (v) development process (items
23–28), (vi) using evidence (items 29–33), (vii)
disclo-sure and transparency (items 34–35), and (viii) plain lan-guage (item 36). Since not all DAs had been evaluated
in scientific studies, we decided to exclude the two items related to the evaluation dimension. Response options for each criteria item were ‘yes’ and ‘no’ (coded as 1 and 0,
respectively). For each DA, the number of IPDAS items met was converted to percentages of the total number of items.
Communicative aspects of decision aids
The use of communicative aspects by the DAs was assessed by a recently developed and validated commu-nicative aspect checklist for patient DA (Supplementary Material 3) [12]. This tumor-independent checklist con-sists of 76 items divided into seven CAs: (i) information
presentation (items 1–26), (ii) information control (items
27–33), (iii) personalization (items 34–40), (iv)
interac-tion (items 41–55), (v) accessibility of informainterac-tion (items
56–64), (vi) suitability of information (65–68), and (vii)
source of information (items 69–76). Response options for
each item were ‘yes’ and ‘no’ (coded as 1 and 0, respec-tively; seven items needed to be recoded). Since six items were only applicable to web-based DA, the total number of items for paper-based DAs was 70, and for web-based 76. For each DA, the number of communicative aspect items met was converted to percentages of the total number of items. Note that a higher communicative aspects score does not necessarily indicate a higher quality DA; it only suggests that more items from the communicative aspects checklist were taken into consideration.
Results
Search results and decision aid characteristics
In total, 8073 records were identified through five databases, and four additional records through other sources (Fig. 1). Screening titles, abstracts, and full-texts yielded ten eli-gible studies, including seven unique DAs. An additional search through online sources resulted in another 14 unique DAs, leading to a total of 21 DAs included in this review (Table 1). Ten aids originated from the United States, five from the Netherlands, five from Australia, and one from Canada. Eleven of the DAs were web based and ten were paper based. Most DAs discussed reconstruction surgery (11) and/or surgery (10; mastectomy vs. breast-conserving therapy) as treatment options, followed by (adjuvant) radio-therapy (9), systemic radio-therapy (7; (neo)adjuvant chemother-apy and hormonal therchemother-apy), and lymph node surgery (3; axil-lary dissection and sentinel node biopsy). Year of last update ranged from 2008 to 2018, but most (13) had been updated in 2017 or 2018. Seven DAs had 1 or more associated stud-ies [21–30] of which three were RCTs, five evaluation and/ or development studies, and two protocol studies (Table 2).
Quality of decision aids
None of the DAs met all of the IPDAS criteria, and the total percentage of IPDAS criteria met by the DAs ranged from
31 to 92% (mean IPDAS score (M) = 64%, standard devia-tion (SD = 20%), see Fig. 2). The seven DAs with associ-ated studies had slightly higher IPDAS scores (M = 68%, SD = 8%) than DAs without associated studies (M = 63%,
Table
1
Summar
y of t
he decision aids included in t
he sy stematic r evie w ID Title Or ganization/aut hors Countr y Tr eatment discussed Las t update For mat 1 Mas tect om y or br eas t-conser ving t her ap y Har wood (2011) AU S BCS; mas tect om y Aug 2008 Paper 2 Axillar
y dissection or a sentinel node biopsy
Har woor d (2011) AU S Lym ph node sur ger y Aug 2008 Paper 3 Unders tanding duct al car
cinoma in situ (DCIS)
and deciding about tr
eatment
National br
eas
t and o
var
ian cancer center
AU S BCS; mas tect om y; r adio ther ap y; hor monal t her ap y 2010 Paper 4 Fr ankl
y speaking about cancer
: br eas t r econs truc -tion Cancer suppor t community U SA Recons truction sur ger y 2012 Paper 5 Sur ger y c hoices f or w omen wit h DCIS or br eas t cancer
National cancer ins
titute U SA BCS; mas tect om y; r econs truction sur ger y; l ym ph node sur ger y No v 2012 Paper 6 A guide f or w omen who ar e consider ing br eas t cancer tr eatment wit h c hemo ther ap y and/or hor monal t her ap y bef or e sur ger y Aus tralia and ne w zealand br eas t cancer tr ial gr oup (ANZBCT G) Zdenk ow ski (2016, 2018) AU S (N eo)adjuv ant c hemo ther ap y; (N eo)adjuv ant hor monal t her ap y Dec 2014 Paper 7 A patc hw or k of lif e: One w oman ’s s tor y. F or women making br eas t cancer tr eatment decisions
Dan L. Duncan com
pr ehensiv e cancer center , Jaba ja-W eiss (2006, 2011) U SA BCS; mas tect om y; r econs truction sur ger y; r adio -ther ap y; hor monal t her ap y; c hemo ther ap y Aug 2015 We b 8 Ear ly -s tag e br eas t cancer : Choosing y our tr eatment Healt h dialog U SA BCS; mas tect om y; r econs truction sur ger y; r adio -ther ap y; l ym ph node sur ger y Jul 2016 Paper 9 iCanDecide Cancer sur
veillance and outcomes r
esear ch team, Univ ersity of Mic hig an, Ha wle y (2017a, 2017b, 2018) U SA BCS; mas tect om y; r econs truction sur ger y; r adio -ther ap y; c hemo ther ap y; hor monal t her ap y 2017 We b 10 Br eas t r econs truction: Is it r ight f or y ou? Healt h dialog U SA Recons truction sur ger y Jul 2017 Paper 11 Keuzehulp bors tk ank er PA TIENT+ NL BCS; mas tect om y; r adio ther ap y Aug 2017 We b 12 Keuzehulp bors trecons tructie PA TIENT+ NL Recons truction sur ger y Aug 2017 We b 13 Br eas t REC ONs
truction decision aid (BRE
-COND A) Br eas t cancer ne tw or k A us tralia, W es tmead br eas t cancer ins titute, Macq uar ie U niv ersity , Sher man (2016) AU S Recons truction sur ger y Oct 2017 We b 14 OPTIONS: what ar e m y op tions f or br eas t cancer treatment? W ong (2011) CA N Radio ther ap y; hor monal t her ap y Oct 2017 Paper 15 Br eas t cancer sur ger y op tions Allina Healt h U SA BCS; mas tect om y; r adio ther ap y 2018 Paper 16 Bors trecons tructie k euzehulp Zor gk euzelab NL Recons truction sur ger y Jan 2018 We b 17 Br eas t cancer : should I ha ve br eas t r econs truction af ter a mas tect om y? Healt hwise U SA Recons truction sur ger y Mar 2018 We b 18 Br eas t cancer : should I ha ve c hemo ther ap y f or ear ly -s tag e br eas t cancer? Healt hwise U SA Chemo ther ap y Mar 2018 We b 19 Br eas t cancer : should I ha ve br eas t-conser ving sur ger y or a mas tect om y f or ear ly -s tag e br eas t cancer? Healt hwise U SA BCS; mas tect om y Mar 2018 We b 20 Bors tk ank er k euzehulp Zor gk euzelab Maas tric ht UMC+ NL BCS; mas tect om y; r econs truction sur ger y; r adio -ther ap y; c hemo ther ap y Oct 2018 We b
SD = 5%). The best performing DAs on the IPDAS checklist were DA12, DA14, and DA20 (Fig. 3).
Most aids showed high performance on the dimensions information about treatment options, clarifying values, disclosure and transparency, and decision guidance. For instance, all DAs (100%) presented the available treatment options, with the majority of them explaining both positive and negative features of the options (95%). All aids asked patients to think about positive and negative features of the options that matter most to them (100%). Mixed perfor-mance was observed for items related to evidence, develop-ment process, and outcome probabilities. For instance, as mentioned by the DA or associated paper, almost all aids were reviewed by doctors (95%), but only half of them were reviewed by (52%) or tested with (57%) patients. Five aids (24%) did not contain any outcome probabilities. Of the aids that did contain probability information, many did not adhere to good practice guidance on communicating essen-tial elements such as providing event rates (57%), keeping the same denominators (29%), reporting time period (43%), or uncertainty (52%). Moreover, only four DAs (19%) reported the update policy and three (14%) discussed the quality of the evidence used. Finally, regarding the dimen-sion of plain language, only five aids (24%) reported accept-able readability levels (e.g., 8th–10th grade (Flesch-Kincaid) reading level).
Communicative aspects of decision aids
A full summary of the results on the assessment of commu-nicative aspects can be found in Supplementary Material 3. The overall percentage of communicative aspect items met by the DAs ranged from 31% to 68% (M = 52%, SD = 10%). The seven DAs with associated studies had similar com-municative aspects scores (M = 52%, SD = 5%) compared to DAs without associated studies (M =52%, SD = 2%). The best performing DAs on the communicative aspects check-list were DA9, DA20, and DA21 (Fig. 3). In general, the majority of the aids met most items related to accessibility; mixed results were found for items with respect to informa-tion presentainforma-tion, informainforma-tion control, interacinforma-tion, and suit-ability of information; the least number of items met was shown for personalization and source of information (Fig. 4).
Information presentation
All DAs used different presentation formats for communicat-ing outcome probabilities. Of the aids, 3 (14%) did not use any method, 2 (10%) used words-only (e.g., verbal descrip-tions), 6 (29%) used a combination of words and numbers, and 10 (47%) applied a combination of words, numbers, and visuals. Of the 16 aids that used numerical methods, natural frequencies were most often used (12; 75%) followed
Table 1 (continued) BCS br eas t-conser ving sur ger y ID Title Or ganization/aut hors Countr y Tr eatment discussed Las t update For mat 21 Bors tk ank er RA dio ther apie S Amen beslissen (BRAS A) MAAS TR O Clinic, Maas tric ht U niv ersity , Ne ther
lands Cancer Ins
titute NL Radio ther ap y Oct 2018 We b
Table 2 Summar y of t he s tudies included in t he sy stematic r evie w DA ID Firs t aut hor , y ear Countr y Study design Study population Me thods Results 1,2 Har wood, 2011 AU S De velopment/Ev
aluation and pilo
t study De velopment/Ev aluation s tudy :
women who had alr
eady had sur ger y f or br eas t cancer (s tag es I and II, n = 28) Pilo t s tudy : ne wl y diagnosed patients wit h ear ly br eas t cancer (s tag es no t mentioned, n = 11) Ther e w er e tw o phases of t his s tudy . The firs t phase in vol ved patients ev aluating t he tw o D As, and t he second phase in vol ved de ter min -ing t he effectiv eness of t he D As. Dur ing bo th phases, s tudy outcomes wer e tr eatment c hosen, patient kno wledg e, decisional conflict, and satisf action wit h decision-making. Patients in t he his tor ical contr ol g roup repor ted positiv e f eedbac k on t he
DAs, and patients in t
he inter ven -tion pilo t g roup f ound t he D As t o be helpful. R esults fr om t he pilo t s tudy sugg es ted a possible r eduction in
decisional conflict, and incr
ease in decisional satisf action, kno wledg e, and c hoice of axillar y clear ance (ins
tead of sentinel node biopsy) in
the inter vention pilo t g roup. 6 Zdenk ow ski, 2016 AU S De velopment/pr ot ocol e valuation study Ne wl
y diagnosed patients wit
h in va -siv e and oper able br eas t cancer (tar ge t n = 50) A pr e-pos
t design will be used t
o ev aluate t he accep tability and feasibility of t he D A . Pr imar y
outcomes will be accep
tability
and f
easibility
, and secondar
y
outcomes will be decision conflict, kno
wledg e, inf or mation and in vol vement pr ef er ence, ag reement be tw een pr ef er red and ac hie ved decision. N. A . 6 Zdenk ow ski, 2018 AU S Ev aluation s tudy (pr e-pos t design) Ne wl
y diagnosed patients wit
h oper able in vasiv e br eas t cancer (n = 59) Patients firs t com ple ted a baseline ques tionnair e (tes t 1), subse -quentl y r eceiv ed t he D A pr ior t o consult ation, and t hen com ple ted a f ollo w-up q ues tionnair e af ter consult ation (tes t 2), bef or e sur ger y (tes t 3) and 12 mont hs af ter r egis tration (tes t 4). S tudy outcomes: as abo ve. The D A w as f ound t o be f easible (wit h mos t patients ha ving accessing
it) and accep
table (wit h t he ma jor -ity of t he patients seeing t he D As as useful f or t
heir decision about
treatment). Mor eo ver , pos t-D A ,
decisional conflict, anxie
ty , and dis tress decr eased significantl y. 7 Jiba ja-W eiss, 2006 U SA Ev aluation s tudy Ne wl
y diagnosed patients wit
h ear ly br eas t cancer (s tag es I–IIIA , n = 51) Patients answ er ed a number of ques tions af
ter diagnosis, and
af ter com ple ting t he D A . S tudy outcomes w er e patients ’ use of the v alues clar ification e xer cise, per ceiv ed clar ity of v alues, and
decision conflict scor
es (lo w liter acy v ersion). Ov er half of t he par ticipants per -for med t he v alues clar ification ex er
cise. The use of t
he D A w as associated wit h lo wer le vels of
decisional conflict (com
par ed t o baseline scor es) and lo wer le vels of
feeling unclear about v
Table 2 (continued) DA ID Firs t aut hor , y ear Countr y Study design Study population Me thods Results 7 Jiba ja-W eiss, 2011 U SA RCT Ne wl
y diagnosed patients wit
h ear ly br eas t cancer (s tag es I–IIIA , n = 76) Patients w er e r andomized t o eit her the inter vention g roup (D A plus usual car e) or t he contr ol g roup (usual car e onl y). S tudy outcomes wer e tr eatment pr ef er ence, br eas t cancer kno wledg e, satisf action wit h decision, satisf action wit h decision-making pr ocess, and
decision conflict (lo
w liter acy version). Patients who r eceiv ed t he D A w er e mor e lik ely t o indicate a pr ef er ence for mas tect om y r at her t han br eas t-conser ving sur ger y, w er e mor e kno wledg
eable and clear
er about their v alues com par ed t o t he contr ol gr oup. N o differ ences w er e f ound in satisf action wit h t he decision or t he decision-making pr ocess be tw een the tw o g roups. 9 Ha wle y, 2016 U SA Ev
aluation and pilo
t s
tudy
Ne
wl
y diagnosed patients wit
h ear ly br eas t cancer (s tag e 0, I, or II, n = 101) Patients w er e r andomized t o eit her the inter vention g roup (who vie wed t he D A firs t) or t he contr ol gr oup (who t ook a sur ve y pr ior t o vie wing t he D A). S tudy outcomes for t he e valuation w er e kno wl -edg e (about tr eatment op tions and br eas
t cancer) and decisional
appr
aisal.
Patients who vie
wed t
he D
A firs
t
had higher scor
es on decisional appr aisal t han t he contr ol g roup. Ho we ver , no s tatis ticall y significant differ ences w er e f ound in kno wledg e about tr eatment op tions be tw een t he tw o g roups. 9 Ha wle y, 2017 U SA RCT pr ot ocol Ne wl
y diagnosed patients wit
h ear ly br eas t cancer (DCIS, or s tag e I–II, tar ge t n = 222 per ar m) A tw o-ar m R CT will be conducted to e valuate t he im pact of a t ailor ed DA (inter vention g roup) on deci -sion q uality , decision satisf ac -tion, deliber
ation, and decision
pr epar edness (as pr imar y s tudy outcomes) com par ed t o t he same non-t ailor ed s tatic D A (contr ol gr oup). N. A . 9 Ha wle y, 2018 U SA RCT Ne wl
y diagnosed patients wit
h ear ly br eas t cancer (s tag e I–II, n = 496) Patients w er e r andoml y allocated to t he inter vention g roup (t ailor ed DA) or contr ol g roup (non-t ailor ed static D A). Pr imar y s tudy outcome was high-q uality decision-making (whic h consis ted of (1) kno wl -edg e about r isk s and benefits of treatment op tions and (2) v alues-concor dant tr eatment). The use of a t ailor ed D A w as posi -tiv
ely associated wit
h high-q uality decisions com par ed t o using a non-tailor ed D A . F ur ther mor e, patients in t he inter vention g
roup had higher
lev els of kno wledg e t han t he contr ol gr oup. Ho we ver , no differ ences wer e f ound in v alues-concor dant treatment decisions be tw een t he tw o gr oups.
Table 2 (continued) DA ID Firs t aut hor , y ear Countr y Study design Study population Me thods Results 13 Sher man, 2016 AU S RCT Ne wl
y diagnosed patients wit
h ear ly br eas t cancer or duct al car ci
-noma in situ (DCIS, s
tag es I–III, n = 222) An R CT w as conducted t o de ter -mine t he effectiv eness of a D A f or deciding whe ther t o ha ve br eas t recons truction or no t. P atients wer e r andomized t o eit her t he inter vention gr oup (D A + plus standar d inf or mation) or t he con -trol g roup (s tandar d inf or mation). Study outcomes w er e decisional conflict, satisf action wit h inf or ma
-tion, and decisional r
eg re t (1 and 6 mont hs af ter e xposur e). At bo th 1- and 6-mont h f ollo w-up, t he use of t he D A w as associated wit h lo wer le
vels of decisional conflict
and higher le vels of satisf action wit h the inf or mation com par ed t o t he contr ol g roup. Ther e w er e no differ -ences in decisional r eg re t be tw een the tw o g roups. 14 W ong, 2011 CA N De velopment and e valuation s tudy De velopment s tudy : patients wit h ear ly br eas
t cancer who had
alr eady had r adio ther ap y (s tag e I, n = 12) Ev aluation s tudy : N ew ly diagnosed patients wit h ear ly br eas t cancer (s tag e I, n = 36) Ther e w er e tw o pilo t s tudies in this s tudy . The firs t in vol ved t he de velopment of t he D A in whic h patients w er e ask ed t o r evie w t he accep tability of t
he aid. The sec
-ond pilo t w as a pr e-pos t tes t aimed at e xamining t he effectiv eness of t he D A on decisional conflict, kno wledg e, im pact of e vent, and treatment c hoice. The ma jor ity of patients r ated t he D A as (e xtr emel y) satisfied. In com par i-son t o t he baseline scor es (pr e-tes t), patients e xper
ienced less decisional
conflict and w er e mor e kno wledg e-able af ter using t he D A (pos t-tes t). DA decision aid, N. A. no t applicable, RCT randomized contr olled tr ial
Fig. 2 The international patient decision aid standard (IPDAS) scores for each decision aid. Decis guidan decision guidance, D&T disclosure and transparency, PL plain language
Fig. 3 Percentage of items met
on the IPDAS and communica-tive aspects checklist for each decision aid. Decision aids are presented in chronological order (based on year of last update)
by percentages (11; 69%); for the 10 aids that used visual methods, icon arrays were the most common (9; 90%), fol-lowed by a pie chart or line graph (both 1; 10%). Of the 18 aids that communicated probability information, 14 (78%) described uncertainties around them, typically with verbal methods (13; 93%), followed by numerical ranges (8; 57%), and visually presented confidence intervals (1; 7%). Varia-tions were also observed in presenting disease-related infor-mation (6 used text-only, 10 a combination of text and vis-ual/audiovisual), and procedures of treatments (6 text-only, 15 a combination of text and visual/audiovisual). Finally, a significant number of DAs (19; 90%) presented informa-tion in an unbalanced way; 9 aids (43%) used more space/ text for a specific treatment option, the majority provided an unequal number of positive (12; 55%) and negative features (17; 85%) across the treatment options, and of the 16 aids that included statistical information only 5 (31%) displayed such statistics in a similar way for each option.
Personalization
The majority of the DAs (14; 67%) were tailored towards the breast cancer stage (e.g., early-stage). However, tailor-ing towards the type of treatment (7; 33%), specific popula-tions (3; 14%), or other breast cancer-related factors (4; 19%) (e.g., HER2 status) occurred less frequently. Five aids (24%) allowed patients to tailor the content of the DA, 3 (14%) to tailor information to patients’ own preference for the mode of information presentation, and only 2 DAs (10%) allowed patients to view individualized outcome probabilities based on their own situation.
Interaction
Several interaction methods had been used by the DAs. For comparing treatment options (20; 95%), most used side-by-side tables or verbal comparisons (both 17; 85%), 6 (30%)
Fig. 4 Violin plots of the
percentage of items met on the communicative aspects check-list separated for each aspect. For each violin plot, dark dots represent the DAs
included ranking or rating exercises, and 2 (10%) applied conjoint analysis/visual analogue scales based on patients’ preferences. For clarifying patients’ values, the majority (20; 95%) passively asked patients to think about their personal values, and about half used active methods such as weight-ing exercises (12; 60%) and/or sliders to assign values to preferences (9; 45%). Feedback was also given in different ways. Twelve aids (57%) showed the progress of the aid, 12 (57%) provided a summary of patients’ values and prefer-ences, 17 (81%) included a print option. About half (10; 48%) provided space for note taking, and 8 (38%) included a knowledge test.
Information control
Nine aids (43%) allowed patients to only receive information that they wanted to read. The majority (18; 86%) provided a step-by-step way to move through the DA, and 16 (76%) gave patients the opportunity to read more about a specific topic of interest. Only 5 aids (24%) allowed for patients to search for specific keywords or topics in the aid.
Accessibility and suitability
Regarding the suitability of information, almost all DAs (19; 90%) used a conversational (writing) style, and only 6 (29%) contained irrelevant illustrations that did not have any link with the messages being presented. Of the aids that included audiovisual material, only 1 (17%) had videos of less than 1 min. Most aids (16; 76%) were lengthy and contained more than ten (web) pages. Regarding accessibility of the aids, 16 (76%) were freely available on the web, and 5 (24%) required a login code to get full access. Thirteen DAs (62%) reported the date of last update, but only 2 (10%) reported the update frequency. All except for 1 aid could be used on multiple devices such as a laptop or smartphone, or were self-administered. Six aids (29%) required staff assistance in order to start with the aid.
Source of information
Of the 18 DAs that communicated outcome probabilities, most included probabilities for treatment side-effects (12; 67%), followed by recurrence of cancer (12; 67%). Numeri-cal information related to survival rates (4; 22%) or quality of life outcomes (5; 28%) occurred less frequently. Only 5 DAs (28%) reported the original source of the probabilities (e.g., RCTs or population-based data), of which 3 (60%) pro-vided detailed information about the patients included in the data (sets) and 1 (20%) about the period of data collection.
Discussion
In this systematic review, we identified 21 currently avail-able patient DAs for early-stage breast cancer treatment, and critically reviewed their quality (as assessed by the IPDAS checklist [10]) and use of communicative aspects (as assessed by a communicative aspect checklist [12]). This review shows substantial variability in the quality of the DAs, with no existing DA meeting all of the inter-nationally agreed IPDAS criteria. Many did not adhere to good practice guidance on providing information about the development, evidence used for the content, or report-ing readability levels. This limited adherence to the qual-ity criteria has also been found among existing DAs for patients with localized prostate cancer [7, 12]. Never-theless, it is promising to see that most of the recently launched or updated DAs in our review (i.e., from 2017 onwards) have shown increased adherence to the IPDAS criteria (see Fig. 3), which suggests that current DA devel-opers and/or clinicians are now taking these criteria much more into account than in the past. At the same time, how-ever, patients can still easily find and make use of existing low-quality DAs, which may foster low implementation rates [5, 6].
We also observed that few DAs presented a thorough description of outcome probabilities of treatment options. In fact, three aids did not contain any probability infor-mation at all, and two only used verbal descriptions. Ide-ally, treatment decision-making is, among other elements such as patients’ preferences, guided by evidence-based probabilities of treatment outcomes such as survival rates, side-effects, or quality of life after treatment [3, 13]. Fol-lowing the IPDAS guidelines, such outcomes may help newly diagnosed cancer patients in balancing the risks and benefits of options together with their clinician, and should therefore be incorporated in DAs [31]. Moreover, from an ethical point of view, patients should be fully and ade-quately informed3, and thus they should also be informed about outcome probabilities and their original sources [32]. The lack of statistical information for breast cancer DAs is remarkable and in contrast with DAs evaluated for men with localized prostate cancer of which all (except for one DA) contained numeric estimates regarding survival rates and side-effects of treatments [12].
The DAs that did communicate probability informa-tion showed great variability in how they communicated such statistical information. Most aids used numeric esti-mates such as natural frequencies or percentages, and only a few used visual aids such as icon arrays. However, several studies have shown that patients (especially with low numeracy skills) often misunderstand such statistics [33], especially when only being communicated in words
[34]. Adding numbers in combination with visual aids may facilitate patients’ understanding of probabilities and overcome several biases such as denominator neglect or framing effects [13]. This multimodal strategy (e.g., using both words and pictures) is also useful for communicat-ing other treatment information (e.g., procedures of treat-ments), which may lead to better information recall by patients [35]. Over the years, several best practices in the communication of evidence-based outcome probabilities have been developed [13, 33], and it is important that DA developers and clinicians who are communicating statisti-cal information to patients are taking these sets of guiding principles into account.
One of the more significant communicative issues found in the reviewed DAs for early breast cancer concerns the lack of personalization. For instance, all (except for two) DAs communicated average outcome probabilities based on statistics of groups of prior patients, which may be difficult to apply to the situation of individual patients [36]. Clinical decision-support tools for explaining chemotherapy survival benefits exist (e.g., Predict-UK), and can already estimate personalized outcomes based on patients’ personal (e.g., age) and disease-related (e.g., tumor stage) characteristics entered by the clinician. However, such tools are often difficult to understand for patients and should always be used in consul-tation with a clinician. We therefore argue that patient DAs can be improved by incorporating patient-friendly versions (or result pages) of such personalized clinical prediction models into existing or novel DAs. However, a prerequisite for personalizing outcomes to individual patients is the avail-ability of robust predictive models based on large amounts of clinical data [37, 38]. Recent technological advances in data science and artificial intelligence in combination with large population-based (e.g., cancer registries) or patient-reported outcome datasets offer promise for the generation of personalized treatment outcomes in DAs [12, 39].
This review further reveals some potential commu-nicative issues of early breast cancer-specific DAs that could hinder their uptake in routine clinical practice. For instance, most aids provided extensive and detailed information about the options. This may be beneficial for patients who prefer detailed information about treatment options, but may discourage patients who do not have the need, time, or capacity for this [40]. Similarly, not all DAs were easily accessible for patients due to, for instance, lim-ited access (i.e., login code), out-datedness of information, or poor findability. These accessibility issues might be bar-riers for especially patients with low literacy skills, who face difficulty in finding, evaluating, and obtaining online health information [41]. Next to that, clinicians may better appreciate the benefit of using and providing DAs to their patients if communicative aspects such as personalization (e.g., individualized treatment outcomes) or interaction
(e.g., value-clarification exercises) are taken into account. Clinicians may wonder how a limited DA can add to their advisory consult and whether a low literacy patient can take advantage of this DA. It is plausible that improving these communicative aspects of DAs will lower the barrier for clinicians to distribute DAs to their patients.
Our review does have some limitations. First, most DAs were identified through online sources compared to the academic literature. Initially, we found 26 DAs with asso-ciated studies, which was comparable with the number of studies found by a related review [1]. In contrast with that review, we needed to have full access to the tools in order to accurately review their quality and communicative aspects. Hence, we could only obtain full access to a minority of those aids found through the academic sources. It should be noted, though, that this distribution of aids found via pub-lished literature or online sources is similar to distributions found in related reviews [7, 12], that used a similar method for identifying and reviewing the characteristics of DAs. Another limitation is that we could not link the IPDAS and communicative aspect scores to various SDM outcomes, mostly because of the lack of data. For instance, it may be that DAs that are personalized (in terms of content, amount of information, or mode of information delivery) are seen as more personally relevant and processed more deeply by patients [42]. The benefit of this in-depth processing is that patients may acquire better knowledge about their options, which makes them better prepared for their next consulta-tion, with more time actively involved in a SDM process [43].
Conclusion
SDM in early breast cancer care requires that patient and clinician are both well-informed about the clinical case and personal situation at hand. DAs have been developed to facilitate this process, but their implementation in routine clinical practice remains low. This review provides insights into the variability among currently available DAs for early breast cancer treatment, and shows that both their quality and use of various communicative aspects can be improved. In addition, even though adherence to the IPDAS checklist is important for ensuring high-quality DAs, our findings sug-gest that DA developers should also seriously consider com-municative aspects that could influence the uptake of DAs in daily practice. Our results do not only have implications for clinicians who are involved in the development and use of DAs for breast cancer treatment, but also for clinicians outside of breast cancer who are facing similar complex and time-consuming clinical counseling scenarios with their patients.
Acknowledgements We would like to thank the authors and/or organi-zations of the included decision aids for providing access to the tools.
Funding The funding was provided by Data Science Center Tilburg
and The Netherlands Organisation for Scientific Research (NWO) (Grant No. 628.001.030)
Compliance with ethical standards
Conflicts of interest The authors declare that they have no conflict of
interest.
Research involving human participants and/or animals This article
does not contain any studies with human participants or animals per-formed by any of the authors.
Informed consent This article does not contain any studies with human
participants performed by any of the authors. Therefore, obtaining informed consent does not apply.
Open Access This article is distributed under the terms of the
Crea-tive Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribu-tion, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
References
1. Zdenkowski N, Butow P, Tesson S, Boyle F (2016) A systematic review of decision aids for patients making a decision about treat-ment for early breast cancer. The Breast 26:31–45
2. Stacey D, Légaré F, Lewis K et al (2017) Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev 4:CD1431
3. Stiggelbout AM, van der Weijden T, Wit MPTD, Frosch D, Legare F, Montori VM, Trevena L, Elwyn G (2012) Shared decision making: really putting patients at the centre of healthcare. BMJ 344:e256
4. Elwyn G, Durand MA, Song J et al (2017) A three-talk model for shared decision making: multistage consultation process. BMJ 359:j4891
5. Elwyn G, Scholl I, Tietbohl C et al (2013) “Many miles to go.” A systematic review of the implementation of patient decision sup-port interventions into routine clinical practice. BMC Med Inform Decis Mak 13:14
6. Lin GA, Halley M, Rendle KAS, Tietbohl C, May SG, Trujillo L, Frosch DL (2013) An effort to spread decision aids in five Califor-nia primary care practices yielded low distribution, highlighting hurdles. Health Aff 32:311–320
7. Adsul P, Wray R, Spradling K, Darwish O, Weaver N, Siddiqui S (2015) Systematic review of decision aids for newly diagnosed patients with prostate cancer making treatment decisions. J Urol 194:1247–1252
8. Elwyn G, O’Connor A, Stacey D, Volk R, Edwards A, Coulter A (2006) Developing a quality criteria framework for patient deci-sion aids: online international Delphi consensus process. BMJ 333:417–419
9. Feibelmann S, Yang TS, Uzogara EE, Sepucha K (2011) What does it take to have sustained use of decision aids? A programme evaluation for the breast cancer initiative. Heal Expect 14:85–95
10. Elwyn G, O’Connor AM, Bennett C et al (2009) Assessing the quality of decision support technologies using the international patient decision aid standards instrument (IPDASi). PLoS ONE 4:e4705
11. Bekker HL (2010) The loss of reason in patient decision aid research: do checklists damage the quality of informed choice interventions? Patient Educ Couns 78:357–364
12. Vromans RD, van Eenbergen MC, Pauws SC, Geleijnse G, van der Poel HG, van de Poll-Franse LV, Krahmer EJ (2019) Com-municative aspects of decision aids for localized prostate cancer treatment—a systematic review. Urol Oncol Semin Orig Investig 37:409–429
13. Trevena LJ, Zikmund-Fisher BJ, Edwards A et al (2013) Present-ing quantitative information about decision outcomes: a risk com-munication primer for patient decision aid developers. BMC Med Inform Decis Mak 13:S7
14. Witteman HO, Scherer LD, Gavaruzzi T et al (2016) Design fea-tures of explicit values clarification methods. Med Decis Mak 36:453–471
15. Feldman-Stewart D, Brundage MD (2004) Challenges for designing and implementing decision aids. Patient Educ Couns 54:265–273
16. Shannon C, Weaver W (1949) The mathematical theory of com-munication. Univ Illinois Press 5:1–131
17. Obeidat R, Finnell DS, Lally RM (2011) Decision aids for surgical treatment of early stage breast cancer: a narrative review of the literature. Patient Educ Couns 85:e311–e321
18. Waljee JF, Rogers MAM, Alderman AK (2007) Decision aids and breast cancer: do they influence choice for surgery and knowledge of treatment options? J Clin Oncol 25:1067–1073
19. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioan-nidis JPA, Clarke M, Devereaux PJ, Kleijnen J, Moher D (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 339:b2700
20. McHugh ML (2012) Interrater reliability: the kappa statistic. Bio-chem Medica 22:276–282
21. Harwood R, Douglas C, Clark D (2011) Decision aids for breast and nodal surgery in patients with early breast cancer: develop-ment and a pilot study. Asia Pac J Clin Oncol 7:114–122 22. Hawley ST, Newman L, Griggs JJ, Kosir MA, Katz SJ (2016)
Evaluating a decision aid for improving decision making in patients with early-stage breast cancer. Patient 9:161–169 23. Hawley ST, Li Y, Jeanpierre LA, Goodell S, Jagsi R, Ward KC,
Sabel MS, Katz SJ (2017) Study protocol: a randomized con-trolled trial of a comprehensive breast cancer treatment patient decision tool (iCanDecide). Contemp Clin Trials Commun 5:123–132
24. Hawley ST, Li Y, An LC et al (2018) Improving breast cancer surgical treatment decision making: the iCanDecide randomized clinical trial. J Clin Oncol 36:659–666
25. Jibaja-Weiss ML, Volk RJ, Friedman LC, Granchi TS, Neff NE, Spann SJ, Robinson EK, Aoki N, Robert Beck J (2006) Prelimi-nary testing of a just-in-time, user-defined values clarification exercise to aid lower literate women in making informed breast cancer treatment decisions. Health Expect 9:218–231
26. Jibaja-Weiss ML, Volk RJ, Granchi TS, Neff NE, Robinson EK, Spann SJ, Aoki N, Friedman LC, Beck JR (2011) Entertainment education for breast cancer surgery decisions: a randomized trial among patients with low health literacy. Patient Educ Couns 84:41–48
27. Sherman KA, Shaw LKE, Winch CJ et al (2016) Reducing deci-sional conflict and enhancing satisfaction with information among women considering breast reconstruction following mastectomy: results from the BRECONDA randomized controlled trial. Plast Reconstr Surg 138:592e–602e
28. Wong J, D’Alimonte L, Angus J, Paszat L, Metcalfe K, Whelan T, Llewellyn-Thomas H, Warner E, Franssen E, Szumacher E (2012) Development of patients’ decision aid for older women with stage I breast cancer considering radiotherapy after lumpectomy. Int J Radiat Oncol Biol Phys 84:30–38
29. Zdenkowski N, Butow P, Hutchings E, Douglas C, Coll JR, Boyle FM (2016) A decision aid for women considering neoadjuvant systemic therapy for operable invasive breast cancer: development and protocol of a phase II evaluation study (ANZ1301 DOMINO). JMIR Res Protoc 5:e88
30. Zdenkowski N, Butow P, Spillane A, Douglas C, Snook K, Jones M, Oldmeadow C, Fewster S, Beckmore C, Boyle FM (2018) Sin-gle-arm longitudinal study to evaluate a decision aid for women offered neoadjuvant systemic therapy for operable breast cancer. JNCCN J Natl Compr Cancer Netw 16:378–385
31. Trevena LJ, Davey HM, Barratt A, Butow P, Caldwell P (2006) A systematic review on communicating with patients about evi-dence. J Eval Clin Pract 12:13–23
32. Engelhardt EG, Pieterse AH, Han PKJ et al (2017) Disclosing the uncertainty associated with prognostic estimates in breast cancer: current practices and patients’ perceptions of uncertainty. Med Decis Mak 37:179–192
33. Gigerenzer G, Gaissmaier W, Kurz-Milcke E, Schwartz LM, Woloshin S (2008) Helping doctors and patients make sense of health statistics. Psychol Sci Public Interes 8:53–96
34. Büchter RB, Fechtelpeter D, Knelangen M, Ehrlich M, Walter-ing A (2014) Words or numbers? CommunicatWalter-ing risk of adverse effects in written consumer health information: a systematic review and meta-analysis. BMC Med Inform Decis Mak. https :// doi.org/10.1186/1472-6947-14-76
35. Mayer RE (2005) Cognitive theory of multimedia learning. In: Mayer RE (ed) Cambridge handbook of multimedia learning. Cambridge University Press, Cambridge, pp 31–48
36. Kent DM, Hayward RA (2007) Limitations of applying sum-mary results of clinical trials to individual patients. JAMA 298:1210–1212
37. Flynn D, Ford GA, Stobbart L, Rodgers H, Murtagh MJ, Thomson RG (2013) A review of decision support, risk communication and patient information tools for thrombolytic treatment in acute stroke: lessons for tool developers. BMC Health Serv Res 13:225 38. Fraenkel L, Fried RF (2010) Individualized medical decision
mak-ing. Arch Intern Med 170:566–569
39. van Stam MA, van der Poel H (2018) The new standard: per-sonalised information about the risks and benefits of treatment strategies for localised prostate cancer. Eur Urol 74:34–36 40. de Rooij BH, Ezendam NPM, Vos MC, Pijnenborg JMA, Boll D,
Kruitwagen RFPM, van de Poll-Franse LV (2019) Patients’ infor-mation coping styles influence the benefit of a survivorship care plan in the ROGY care trial: new insights for tailored delivery. Cancer 125:788–797
41. Diviani N, Van Den Putte B, Giani S, Van Weert JCM (2015) Low health literacy and evaluation of online health information: a systematic review of the literature. J Med Internet Res. https :// doi.org/10.2196/jmir.4018
42. Lustria MLA, Cortese J, Gerend MA, Schmitt K, Kung YM, Mclaughlin C (2016) A model of tailoring effects: a randomized controlled trial examining the meachanisms of tailoring in a web-based STD screening intervention. Heal Psychol 35:1214–1224 43. Ritterband LM, Thorndike FP, Cox DJ, Kovatchev BP,
Gonder-Frederick LA (2009) A behavior change model for internet inter-ventions. Ann Behav Med 38:18–27
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Affiliations
Ruben Vromans1 · Kim Tenfelde1 · Steffen Pauws1,2 · Mies van Eenbergen3 · Ingeborg Mares‑Engelberts4,5 ·
Galina Velikova6 · Lonneke van de Poll‑Franse3,7,8 · Emiel Krahmer1 Kim Tenfelde
k.tenfelde@uvt.nl Steffen Pauws s.c.pauws@uvt.nl Mies van Eenbergen m.vaneenbergen@iknl.nl Ingeborg Mares-Engelberts i.engelberts@erasmusmc.nl Galina Velikova
g.velikova@leeds.ac.uk Lonneke van de Poll-Franse l.vd.poll@nki.nl
Emiel Krahmer e.j.krahmer@uvt.nl
1 Department of Communication and Cognition,
Tilburg University, Warandelaan 2, 5035 AB Tilburg, The Netherlands
2 Chronic Disease Management, Philips Research, Eindhoven,
The Netherlands
3 Department of Research, Netherlands Comprehensive Cancer
Organization (IKNL), Utrecht, The Netherlands
4 Department of Medical Ethics and Philosophy of Medicine,
Erasmus MC, Rotterdam, The Netherlands
5 Department of Surgery, Sint Franciscus Vlietland Group,
Rotterdam, The Netherlands
6 Leeds Institute of Medical Research at St Jame’s, University
of Leeds, Leeds, UK
7 Division of Psychosocial Research & Epidemiology, The
Netherlands Cancer Institute, Amsterdam, The Netherlands
8 Department of Medical and Clinical Psychology, Tilburg
